A benchmark test for room air distribution: the backward facing step flow

This paper concerns the use of Computational Fluid Dynamics (CFD) for the prediction of room air movement and provide a guide on selecting of the proper turbulence model. The benchmark which developed here for the first time is a backward-facing step flow problem. The measurements is performed in a small-scale model and the velocity filed is measured by Particle Image Velocimetry (PIV). The measurements focus on transitional flow and fully developed turbulent flow at isothermal condition. The PIV measurements and different CFD predictions are compared. The CFD predictions are generally slightly diverse, and this is, among other things, the result of using different turbulence models as well as using different software codes, grid distribution and boundary values etc. The k-ϵ family of turbulence models is an option for fully developed flow, and the k-ω family does work for transitional low turbulent flow in this backward-facing step flow.